Ovarian cancer causes more deaths than any other cancer of the female reproductive system. An estimated 25,580 new cases will be diagnosed during 2004 in the United States, and approximately 16,090 of these women will die of the disease. Despite advances in the management of advanced ovarian cancer, 70% to 80% of patients will ultimately succumb to disease that is diagnosed in late stages. When ovarian cancer is diagnosed in stage I, more than 90% of patients can be cured with conventional surgery and chemotherapy. At present, however, only 25% of ovarian cancers are detected in stage I. Detection of a greater fraction of ovarian cancers at an early stage might significantly affect survival. A worldwide research effort, aiming at early detection of ovarian cancer, is currently being performed; finding molecular markers for the disease is one of the major research topics (J Clin Oncol. 2003 May 15; 21(10 Suppl):200-5).
No single marker has been shown to be sufficiently sensitive or specific to contribute to the diagnosis of ovarian cancer. The marker that is currently most frequently used is CA-125 (Br J Cancer. 2000 May; 82(9):1535-8). Its properties do not support its use for screening, but it is a major diagnostic tool. CA-125 is a member of the epithelial sialomucins markers group and is the most well documented and the best performing single marker from this group. Another name for CA-125 is mucin 16, and although it is a membrane protein, it can be found in the serum. Its greatest sensitivity is achieved for serous and emdometrioid ovarian tumors compared to mucinous or clear cell tumors. Other than diagnosis, it can be used for disease monitoring (Eur J Gynaecol Oncol. 2000; 21(1):64-9). In about 70% of patients, a rising level of CA-125 may be the first indication of relapse, predating clinical relapse by a median of 4 months. The serum concentration of CA-125 is elevated by the vascular invasion, tissue destruction and inflammation associated with malignant disease and is elevated in over 90% of those women with advanced ovarian cancer. Yet, CA-125 is not specific to ovarian cancer. It is elevated in 40% of all patients with advanced intra-abdominal malignancy. Levels can also be elevated during menstruation or pregnancy and in other benign conditions such as endometriosis, peritonitis or cirrhosis, particularly with ascites. CA-125 is not a marker that can be detected through use of urine samples due to a high molecular weight.
There are other ovarian cancer markers originating from epithelial mucins but none can replace CA-125, due to poorer specificity and sensitivity. These other markers may prove complementary to CA-125. CA-50, CA 54-61, CA-195 and CA 19-9 all appear to have greater sensitivity for detection of mucinous tumors while STN and TAG-72 have better sensitivity for detection of clear cell tumors (Dis Markers. 2004; 20(2):53-70).
Kallikreins, a family of serine proteases, and other protease-related proteins are also potential markers for ovarian cancer. Indeed, the entire family of kallikreins map to a region on chromosome 19q which is shown to be amplified in ovarian cancers. In particular, kallikrein 6 (protease M) and kallilrein 10 have been reported to have sensitivity up to 75% and specificity up to 100%. Matrix metalloproteinases (MMPs) are another family of proteases useful in ovarian cancer screening and prognosis. MMP-2 was reported to have 66% sensitivity and 100% specificity in one study. Cathepsin L, a cystein protease, was described to have a lower false positive rate compared with CA-125. Based on their biochemical proteolytic role, it would seem likely that these proteases would be active in invasion and metastasis formation and indeed these markers appear to have higher sensitivity for advanced stages of the disease. Due to their relatively low molecular weight, such proteases are candidates to be urine markers, or markers which can be detected in urine samples (Dis Markers. 2004; 20(2):53-70).
Hormones have a role in normal ovarian physiology. Therefore, it is not surprising that hormones, and growth and inhibition factors as well, are suitable for ovarian cancer detection. Measurements of fragments of gonadotropin in the urine were found to have sensitivity up to 83% and specificity up to 92% for detecting ovarian cancer. Inhibins, members of the Transforming Growth Factors (TGF) beta superfamily, have been shown to have a diagnostic value in the detection of granulosa cell tumor, a relatively uncommon type of ovarian cancer, associated with better prognosis overall. Serum inhibin is an ovarian product which decreases to non detectable levels after menopause, however, certain ovarian cancers (mucinous carcinomas and sex cord stromal tumours such as granulosa cell tumours) continue to produce inhibin. Studies have shown that that inhibin assays which detect all inhibin forms (as opposed to test detecting specific members of the inhibins family) provide the highest sensitivity/specificity characteristics as an ovarian cancer diagnostic test (Mol Cell Endocrinol. 2002 May 31; 191(1):97-103). Measurement of serum TGF-alpha itself was found to have sensitivity up to 70% and specificity of 89% in early stage disease. The growth factor Mesothelin was also found to have diagnostic value but only for late stage disease.
Immunohistochemistry is frequently used to assess the origin of tumor and staging when a pathological tissue sample is available. A few molecular markers have been shown to have diagnostic value in Immunohistochemistry of ovarian cancer, among them Epidermal Growth Factor, p53 and HER-2. P53 expression is much lower at early stage than late stage disease. P53 high expression is more typical or characteristic of invasive serous tumors than of mucinous tumors. No benign tumors are stained with P53. HER-2 is found in less than 25% of newly diagnosed ovarian cancers. Ovarian cancer of type granulosa cell tumor has in general better prognosis with late relapse and/or metastasis formation. However, about 50% of patients still die within 20 years of diagnosis. In this specific tumor type, immunohistochemistry staining of estrogen receptor beta (ERb) and proliferating cell nuclear antigen (PCNA) showed that loss of ERb expression and high PCNA expression, characterized a subgroup of granulosa cell tumors with a worse outcome (Histopathology. 2003 September; 43(3):254-62). Survivin expression was also shown to be correlated to tumor grade, histologic type and mutant p53 but actual correlation to survival is questionable (Mod Pathol. 2004 February; 17(2):264)
Many other markers have been tested over the years for ovarian cancer detection. Some markers have shown only limited value while others are still under investigation. Among them are TPA and TPS, two cytokeratins whose inclusion in a panel with CA-125 resulted in diagnoses with sensitivity up to 93% and specificity up to 98%. LPA—lysophosphatidic acid—was a very promising marker with one study demonstrating 98% sensitivity and 90% specificity. However, this marker is very unstable and requires quick processing and freezing of plasma, and therefore has limited usage.
As previously described, no single marker has been shown to be sufficiently sensitive or specific to contribute to the diagnosis of ovarian cancer. Therefore combinations of markers in panel are being tested. Usually CA-125 is one of the panel members. The best performing panel combinations so far have been CA-125 with CA 15-3 with sensitivity of 93% and specificity of 93%, CA-125 with CEA (which has very little sensitivity by itself) with specificity of 93% and specificity of 93%, and CA-125 with TAG-72 and CA 15-3 where specificity becomes 95% but sensitivity is diminished (Dis Markers. 2004; 20(2):53-70).